Prior aquatic animal image classification research focused on distinguishing external features in controlled settings, utilizing either digital cameras or webcams. Identifying visually similar species, like Short mackerel (Rastrelliger brachysoma) and Indian mackerel (Rastrelliger kanagurta), is challenging without specialized knowledge. However, advancements in computer technology have paved the way for leveraging machine learning and deep learning systems to address such challenges. In this study, transfer learning techniques were employed, utilizing established pre-trained models such as ResNet50, Xception, InceptionV3, VGG19, VGG16, and MobileNetV3Small. These models were applied to differentiate between the two species using raw images captured by a smartphone under uncontrolled conditions. The core architecture of the pre-trained models remained unchanged, except for the removal of the final fully connected layer. Instead, a global average pooling layer and two dense layers were appended at the end, comprising 1024 units and by a single unit, respectively. To mitigate overfitting concerns, early stopping was implemented. The results revealed that, among the models assessed, the Xception model exhibited the most promising predictive performance. It achieved the highest average accuracy levels of 0.849 and 0.754 during training and validation, surpassing the other models. Furthermore, fine-tuning the Xception model by extending the number of epochs yielded more impressive outcomes. After 30 epochs of fine-tuning, the Xception model demonstrated optimal performance, reaching an accuracy of 0.843 and displaying a 11.508% improvement in predictions compared to the model without fine-tuning. These findings highlight the efficacy of transfer learning, particularly with the Xception model, in accurately distinguishing visually similar aquatic species using smartphone-captured images, even in uncontrolled conditions.

The Krabi mouth-brooding fighting fish, Betta simplex Kottelat, 1994, is a critically endangered and endemic fish species in Krabi province, Southern Thailand. Little information is available on its reproductive ecology and early developmental morphology, which are essential for studying its conservation. Generally, B. simplex is considered an adaptable animal that can tolerate lower alkalinity and higher hardness compared to its natural environment conditions. In this study, wild broodstocks of B. simplex were collected from the reported type localities and bred in captivity under laboratory conditions for size-series collection. Some biological aspects of B. simplex in its natural environmental conditions were determined. We found that its flaring and mating behavior was similar to those of bubble-nesting fighting fish but did not involve bubble-nest building. The fertilized eggs and pre-flexion larvae were nurtured in the mouth cavity of parental males within 11–12 (mode = 11) days after fertilization (DAF). The first-release offspring developed to the post-flexion stage with a body size of 4.39 ± 0.01 mm of standard length (SL; n = 6) and then to the juvenile stage within 30 days after release with 11.72 ± 0.62 mm SL (n = 4). Thus, we propose the following linear regression equation for growth prediction by age (DAF) and body size (SL; mm): age = 0.2425 SL + 1.7036 (r² = 0.9549). The findings of this study will deepen our knowledge of the reproduction and ontogeny of B. simplex and contribute to its future conservation and management.

Morphological descriptions of 13 wild fighting fish species were revised by examining specimens collected as near as possible to the type specimen localities. The description has been revised for certainty and consistency. Principal Component Analysis (PCA) and Kruskal–Wallis with Dunn’s post hoc tests were analyzed and revealed previously described morphological characteristics were evaluated for distinguishing features. The PCA showed that the high loading factor of the principal component 1 (PC1) indicated that head length (0.221) was beneficial for the fish species identification. The Kruskal–Wallis test indicated the significant differences in morphometric characteristics among the fish species (p < 0.05). Then the significant differences in Dunn’s post hoc test (p < 0.001) in head length distinguished the mouth-brooders (i.e., Betta apollon, B. ferox, B. kuehnei, B. pallida, B. pi, B. prima, B. pugnax, and B. simplex) from bubble-nesters (i.e., B. imbellis, B. mahachaiensis, B. siamorientalis, B. smaragdina, and B. splendens) with a head length range of 29.27–36.42 vs. 24.14–28.31% of standard length, including the symmetrical caudal fin. Some characteristics distinguish the fish into subgroups and some species, such as a greater number of anal soft fin rays and lateral line scales of B. pi among the mouth-brooding species (28–29 vs. 22–26 fin rays and 33–35 vs. 26–32 scales) or of B. smaragdina among the bubble-nesting species (26–28 vs. 22–25 fin rays and 31–33 vs. 27–30 scales). Additionally, descriptive characteristics such as anal and caudal fin shape and melanophore pigmentation on the head, body, and fins, including the coloration of live specimens, are useful for distinguish among species and generating a dichotomous identification key.

The family Sciaenidae comprises around 300 species and 70 genera, with the Indo-West Pacific (IWP) region exhibiting the highest species richness. The genus Boesemania Trewavas, 1977 was established as a monotypic group based on the swim bladder and otolith morphology of Johnius microlepis Bleeker. Corvina polykladiskos Bleeker was recognized as the third species of the genus Bahaba. Based on recent morphology and molecular evidence, we assign Bahaba polykladiskos to the genus Boesemania. Boesemania polykladiskos is distinguished from Boesemania microlepis by its distinctly larger scales, the number of lateral-line, upper and lower trans-lateral-line rows, and sagitta and swim-bladder morphology. Phylogenetic analysis of 24 sciaenid species representing 23 genera from the Indo-West Pacific region further confirms that Boesemania is monophyletic and that Boesemania polykladiskos is the sister species of Boesemania microlepis. The two Boesemania species are commercially important fish in their native range and play an essential role in the global luxury dried swim-bladder trade. However, significant knowledge gaps persist regarding their ecology, biology, and population status. Consequently, this paper also addresses the need for further research and conservation measures for these species.

Within the family Osphronemidae, mouth-brooding fighting fishes are small, air-breathing fish species that hold their eggs and offspring in their buccal cavities as a parental care behavior, usually found in running waters and distributed in the wild throughout Southeast Asia. This study aims to describe the morphological development and generate identification keys for the larval and juvenile stages of seven wild mouth-brooding fighting fish species found in Thailand, i.e., Betta apollon , B. ferox , B. pallida , B. pi , B. prima , B. pugnax , and B. simplex from the wild. The broodstocks were collected from type localities or based on characteristics that most closely matched each species description, with healthy fishes selected and breeding continued for our size-series collection, including B. simplex , a previous study. The results showed that mouth-brooders released their offspring when developed to the post-flexion stage within 11–12 (mode = 11) days after fertilization (DAF), except B. pi , which took 18–20 DAF, and the post-flexion larva developed to the juvenile stage within 18–30 days after release. The main characteristics of the new-release post-flexion larva were an oblong and depressed body, a large head, an oval to rounded eye, and rays where the caudal fin began to develop, as well as having fully developed ventral fins with two or three dorsal, central, and ventral stripes and a caudal spot. Myomere numbers and fin rays differed among species across a range of 8–10 dorsal, 10–13 pectoral, 6 ventral, 24–31 anal, and 10–13 caudal fin rays. Diagnostic characters were selected to create a dichotomous identification key, with an illustration provided. In terms of taxonomy, the different larval stages also differed in pigmentation patterns among species, with pigmentation patterns on head (pre-orbital, sub-orbital, post-orbital, and sub-opercular bands) and longitudinal stripes on the side of the body able to be used to distinguish among different development stages as well as different species.